Table 3. Number of significantly selective neurons in each condition.

Listed are the counts resulting from an ANOVA run on each neuron across trials, with factors number of drops of reward and flavor of reward. The χ² test listed in the third row of each section compares the proportion of neurons in each category with what would be expected by chance. For the p-values with asterisks, this test revealed that the overall number of significant neurons was not more than would be expected by chance; for the other p-values, this test revealed whether the proportion of big-preferring vs. small-preferring or chocolate-preferring vs. vanilla-preferring was significantly biased away from 50/50. In addition to the tests shown, other χ² tests revealed that in both shams and lesions, neurons were more likely to be significant late in blocks in the preferred vs. the anti-preferred directions (p<0.0001 and p<0.001, respectively), and that in the preferred direction late in blocks, lesions had a significantly different distribution than would be expected based on the sham distribution (p<0.0001).

		preferred direction		anti-preferred direction
		early in blocks	late in blocks	early in blocks	late in blocks
sham	big-preferring	1	22	4	3
	small-preferring	4	3	0	8
	χ2 test p	0.34*	<0.001	0.69*	0.13
lesion	big-preferring	0	4	0	4
	small-preferring	2	4	2	3
	χ2 test p	0.53*	1.0	0.53*	0.71
sham	choc-preferring	8	10	2	1
	van-preferring	1	5	2	1
	χ2 test p	<0.05	0.20	0.69*	0.46*
lesion	choc-preferring	3	6	1	3
	van-preferring	1	7	1	3
	χ2 test p	0.32	0.78	0.53*	1.0